U.S. Natural Disasters Exploratory Analysis

Authors

Isaac Ajeto (ajeto@uw.edu)
Raymond Smith (smithray@uw.edu)
Angelina Li (yuchel9@uw.edu)
Amisha G. (amishg@uw.edu)

Date

Spring 2022
May 14th, 2022

Abstract

We analysed U.S. Natural Disaster Declarations since 1953. Our main idea was to tie the occurrence of natural disasters in the U.S. as an effect of climate change over the last 68 years.

Keywords

Keywords: natural disasters; united states; climate change;

Introduction

Briefly introduce your project. Include 3-5 research questions. What motivates the questions? Why are they important? (at least 200 words)

For this project, we will be looking at U.S. Natural Disaster Declarations since 1953. We aim to correlate our findings with climate change so that we can make our readers aware of the harmful effects of climate change. Through this project we aim to answer the following questions:

In which year did most disasters occur?

This question throws light on the year during which most natural disasters occurred. Answer to this question would help us in possibly focusing on certain features of climate change that were most prominent in that year to strengthen the correlation between climate change and natural disasters.
What is the change in the amount of natural disasters in Washington State?

This question helps focus on the change in natural disasters in Washington State. Focusing on a small area can help us understand and correlate the change in natural disasters with features of climate change that are specific to Washington State (such as an astronomical increase in mean temperature),
Which U.S. state experienced the greatest increase in natural disasters since 2000?

This question throws light on the states that are most prone to a certain category of natural disasters. Answer to this question can help disaster managers to increase their presence in certain areas that may require more attention so that rescue operations can be conducted in a timely manner.

Related Work

The growing occurrence of natural disasters in the United States is one of the most pressing consequences of climate change. According to the United States Geological Survey, “With increasing global surface temperatures the possibility of more droughts and increased intensity of storms will likely occur. As more water vapor is evaporated into the atmosphere it becomes fuel for more powerful storms to develop. More heat in the atmosphere and warmer ocean surface temperatures can lead to increased wind speeds in tropical storms. Rising sea levels expose higher locations not usually subjected to the power of the sea and to the erosive forces of waves and currents” (USGS, 2021).

With the increasing prevalence of natural disasters such as hurricanes, wildfires, or tornados, being able to use aggregate data to visualize the trends and changes in natural disasters will offer insights on our changing world. This sentiment is ubiquitous across much of the scientific world, and there is a bulk of research on the changing nature and frequency of natural disasters. The World Meteorological Organization found in their 2021 report that natural disasters have increased five-fold over the last 50 years, showing that there is a changing trend in the amount of natural disasters (WMO, 2021). NPR has found that not only are disasters becoming more severe, they also have dire effects on our mental health (Pruitt-Young, 2021). A related work to our project is the Natural Disaster article by Hannah Ritchie and Max Roser, which displays natural disaster deaths over time by country and type (Ritchie, 2021). However, we’d like to focus specifically on FEMA declarations in the USA and the frequency of the disasters themselves, not the amount of deaths. This project will fit into this topic and space by supporting findings of increasing natural disasters with a data visualization. Although related works have focused on deaths, we will focus more specifically on the change in frequency and concentration of natural disasters within the United States, offering insights for those in the United States.

The Dataset

We found our data on kaggle.com, a website created by machine learning practitioners and data scientists. A link to the data will be included at the end of this section. The data was originally collected by the Federal Emergency Management Agency, FEMA. The data set was taken in by Kaggle, which cleaned up the data and formatted it to make it more friendly for use. As stated, this data was collected by FEMA, and the data from 1998 to the present comes from the National Emergency Management Information System (NEMIS). Data from before the launch of the system was manually entered from historical accounts and data regarding previous natural disasters. The data was collected so that FEMA could fulfill their purpose and responsibility of protecting people before, during and after disaster strikes. Our dataset has 4,657 observations, and has 10 columns.

The main ethical question that comes to mind from this dataset has to do with what conclusions are being drawn from it, and if proactive measures are being taken by potentially affected citizens who may be struck by temporary or permanent disasters. The potential limitations of the dataset include the relative novelty of the system. The data collected before 1998 was manually entered, and was derived from historical accounts. The problem with this was that the dataset begins in 1953, and well over half of the 69 years for which we have data were collected from before the NEMIS system was established. As a result of this, there may be some disaster data missing due to the failure of accounts being made for a disaster, or the imperfections of systems to document those disasters. This is not to say that today’s methods are perfect, but technology and information systems have developed so drastically over such a long period of time that it must be acknowledged that the differences in yesterday’s systems may affect the accuracy of the data that we are working with. Another potential problem with the set is that there are so many observations (well over 4,000), which could cause problems to arise while we are attempting to open the set in github. Overall, I would say that the data seems pretty trustworthy considering that it comes from a federal agency, but even such a dataset has its limitations.

Implications

Finding out the most common categories of natural disasters can help researchers monitor which kinds of disasters need more frequent and urgent care as well as make the readers aware of the consequences of climate change. This also shows what kind of resources need to be prepared to respond when natural disasters happen. Having the adequate amount would prevent wasteful behavior, but also ensures the safety of people in the area. Knowing about the average time between a particular disaster can guide disaster managers and researchers to better have an evacuation plan and urge change makers to act on climate change. We would also look into which U.S states had the most disasters by category. Answering this question would guide disaster managers and policymakers to increase their attention in certain areas that need to be more cautious.

Limitations & Challenges

One limitation is that the dataset shows previous data of natural disasters in the United States. However, natural phenomena are not predictable. It is difficult for disaster managers to accurately predict future natural disasters. Although they could have a big idea, there are still flaws to the prediction. The second limitation is that the dataset only includes natural disasters with broad categories. Natural disasters include many specific categories and this dataset might not be able to consider those. Policymakers and disaster managers might not be very accurate when using data and taking measures to help specific areas with a particular natural disaster. The data retrieved from the technology might not be the best to predict future natural disasters. There are limitations to the accuracy of technology. This would be a challenge when disaster managers are considering the next move for each area. These are the challenges and limitations that might affect the result of the data.

Summary Information

Write a summary paragraph of findings that includes the 5 values calculated from your summary information R script

Table

This table displays the number of disasters that occurred within every 5 year period of our dataset. It also includes the number of the 6 most common categories of disasters and the state with the most disasters within that time frame. We included this information because we wanted to make it easier for viewers to see the change over time in disasters. Because our initial data uses each disasters as a new observation, it’s difficult to understand or visualize how many disasters occurred over a year, range of years, or overall. By aggregating this data, we are able to display the change over time in disasters.

Aggregate Data on United States Disaster Declarations in 5 Year Ranges
Years	Number of Disasters	Flood	Hurricane	Tornado	Earthquake	Fire	Severe Storm(s)	State With Most Disasters
(1950,1955]	44	17	14	8	1	1	0	MA
(1950,1955]	44	17	14	8	1	1	0	TX
(1955,1960]	62	34	8	7	1	2	5	OK
(1960,1965]	615	362	62	112	7	1	8	IA
(1965,1970]	1132	826	114	145	0	18	3	MN
(1970,1975]	2670	1957	31	284	2	16	294	PR
(1975,1980]	2908	773	122	105	2	26	174	MI
(1980,1985]	854	414	31	108	4	31	162	CA
(1985,1990]	1729	871	186	94	14	28	447	TX
(1990,1995]	3954	1125	112	241	11	56	1063	TX
(1995,2000]	6458	975	1130	97	1	875	2506	TX
(2000,2005]	10545	270	4610	115	39	502	3180	TX
(2005,2010]	7862	277	731	20	10	932	4531	TX
(2010,2015]	6353	973	1302	48	13	543	2927	MO
(2015,2020]	15201	1516	3166	80	122	436	1651	GA
(2020,2025]	2607	120	871	107	0	123	339	LA

Chart 1

This chart shows the trend of natural disasters from 1953 to present. As we can see, there is an upward trend line from 1992 with a peak in 2020 with a record 9489 natural disasters. There is also a noteworthy decline in disasters after 2020, possibly due to low human activity. This chart helps us correlate the increase in the number of natural disasters with climate change in terms of time.

Chart 2

This chart shows the number of natural disasters that happened in Washington State from 1953 to now. From the graph, the number of natural disasters increased and peaked in 2020. The percent change in the amount of natural disaster would be 136%.

## [1] 136

Chart 3

Describe why you chose this chart and what information it reveals

Citations

“How Can Climate Change Affect Natural Disasters?” How Can Climate Change Affect Natural Disasters? | U.S. Geological Survey, https://www.usgs.gov/faqs/how-can-climate-change-affect-natural-disasters#:~:text=With%20increasing%20global%20surface%20temperatures,more%20powerful%20storms%20to%20develop
“Climate and Weather Related Disasters Surge Five-Fold over 50 Years, but Early Warnings Save Lives - WMO Report | | UN News.” United Nations, United Nations, https://news.un.org/en/story/2021/09/1098662.
Ritchie, Hannah, and Max Roser. “Natural Disasters.” Our World in Data, 3 June 2014, https://ourworldindata.org/natural-disasters.
Pruitt-Young, Sharon. “Climate Change Is Making Natural Disasters Worse - along with Our Mental Health.” NPR, NPR, 11 Sept. 2021, https://www.npr.org/2021/09/11/1035241392/climate-change-disasters-mental-health-anxiety-eco-grief.